 | | | high context switching and high load averages slowing down sy stem | high context switching and high load averages slowing down sy stem 2005-07-28 - By Tom Sightler
Back
On Thu, 2005-07-28 at 16:51 -0400, Magnus Andersen wrote:
> Here is some information from vmstat, iostat, and dstat taken during
> the last peak when I reached a load average ogf 13.
>
> vmstat 2 output:
> procs memory swap io system cpu
> r b swpd free buff cache si so bi bo in cs us sy wa id
> 3 0 12156 21944 125688 2945324 0 0 20 178 368 289556 1 39 0
59
> 8 0 12156 17740 125756 2947200 0 0 1060 184 579 284958 5 38 2
55
> 5 0 12156 22316 125828 2944948 0 0 5798 190 315 276098 6 39 1
54
> 3 0 12156 23200 125932 2945644 0 0 146 1154 576 288639 3 41 1
55
> 3 0 12156 21584 125968 2945644 0 0 0 78 176 303064 1 38 0
61
Wow, that is a lot of context switches. There is also a high amount of
system time with very little user time, that would imply that this is
not an issue with high Oracle usage, unless it's a problem with the SGA
and shared memory. Any information on what waits your Oracle system is
reporting during these low performance periods that might give us a
little clue.
My best guess is something else external, like a driver. Any
possibility of a huge amount of network traffic during this time? Also,
can you post which modules you have loaded and the output of "top -b >
top.txt" during the event would be interesting.
Does the system have periods of relative idle? What do your context
switches look like during that period?
Later,
Tom
--
Taroon-list mailing list
Taroon-list@(protected)
http://www.redhat.com/mailman/listinfo/taroon-list
Earn $52 per hosting referral at Lunarpages.
|
|
|